Enterprise Federation: Essential Research Needed for the GIG Sekar - - PDF document

enterprise federation essential research needed for the
SMART_READER_LITE
LIVE PREVIEW

Enterprise Federation: Essential Research Needed for the GIG Sekar - - PDF document

Enterprise Federation: Essential Research Needed for the GIG Sekar Chandersekaran AF CIO Office and IDA Terry Mayfield IDA August 23 2006 What is the Problem? -1 Distributed systems Spread across multiple enterprises that need


slide-1
SLIDE 1

1

Enterprise Federation: Essential Research Needed for the GIG

Sekar Chandersekaran AF CIO Office and IDA Terry Mayfield IDA August 23 2006

What is the Problem? -1

  • Distributed systems

– Spread across multiple enterprises that need to collaborate tightly to achieve mission objectives – Enterprises [within DOD and across Government organizations and other COI ‘countries ’] are autonomous and make their own choices contributing to heterogeneity – Operational environments dictate heterogeneity

  • Tactical Environment and integration

– Many other factors contributing to heterogeneity

  • Increasing number of protocols
  • Increasingly complex trust relationships
  • Increasing complexity of discovery due to desired ‘DYNAMIC BEHAVIOR’
  • Increasing numbers and types of directories
  • Increasing number of content formats and semantics
  • Business needs of commercial products dictate that they distinguish themselves

based on specialized capabilities

– IM across AOL or Microsoft – Search Engines [Google, Microsoft, Metacrawler, Altavista]

  • Government’s reliance on COTS products and COTS App Dev Environments and

the ‘maxim ’ of no single vendor dependency

slide-2
SLIDE 2

2

What is the Problem? - 2

  • Distributed systems

– A single solution even if it were a universally accepted standard will not suffice

  • POSIX, Linux

– Even within standards there are multiple options that need to be met

  • Profiling is inadequate
  • Dynamic ‘Negotiation’ is needed

– Peripheral IA aspects

  • Systems running in more hostile environments
  • Systems being subjected to more systematic attacks

– Conclusion Dramatically more complex

  • Need to develop new understanding on how to

architect, engineer, manage, and operate.

– Multi Enterprise-Level distributed systems with heterogeneity and diversity using “Federation”

What is Federation?

  • What is federation?

– A federation (Latin: foedus, covenant) is a union comprised of a number of partially self-governing states or regions united by a central ("federal") government. In a federation, the self-governing status of the component states are typically constitutionally entrenched and may not be altered by a unilateral decision of the central government.

  • European Banking Federation, EU

– Application to ‘computing capabilities’

  • WS-Federation (from BEA, IBM, Microsoft, RSA Security, and Verisign,

July 2003) "defines mechanisms that are used to enable identity, account, attribute, authentication, and authorization federation across different trust realms

  • The mechanisms can be used by passive and active requestors; the

Web service requestors are assumed to understand the new security mechanisms and be capable of interacting with Web service providers

  • Ability to integrate in a smooth fashion diverse and heterogeneous but

similar capabilities

– Contributing to ease of use for naïve, power and expert users – Contributing to less complexity in applications – Add complexity to administrators and admin. programs

slide-3
SLIDE 3

3

Fundamental Netcentricity Paradigm

  • SOA all interactions via

‘services’

– Everything modeled as a Service

  • Netcentricity

– Any Consumer to Any Provider

  • User – User or Service
  • Service – User or Service
  • Interactions enterprise wide or

cross enterprise

  • Basic interaction paradigm

– Discover – Select and Locate – Negotiate – Connect – Authenticate – Access

Network Forrest / Enterprise 1

Forest / Enterprise 1 Forest / Enterprise 2 User User Service Service Service User

Interface / Standards

Use Cases

  • WORLDWIDE Cross ENTERPRISE MODEL
  • Each ellipse is a forest
  • Single colored ellipse
  • ne enterprise
  • Each enterprise consists of a number of forests
  • Ellipses of same color are different

forests of the same enterprise

  • Enterprise Trust only between forests
slide-4
SLIDE 4

4

Enterprise Interaction Complexity

Network

Forest / Enterprise 2

User User Discovery Service Service Discovery Service Discovery Expert User

Interface / Standards

Forest / Enterprise 1

Layered Architecture [Large Grain]

Distributed OS [Requestor or Provider], Dist Files, Networking protocol stack Local system management [health, performance, config.], collaboration, messaging Local security, crypto / certificate, time services Web Server Service Provider 1 Data Base Mail etc. SP2 Security SP3 Directory SP4 App Server SP ..n Dist System Capabilities Traditional Middleware Corba or Pre web services [Web] Services Lowest Layer [Web] Services Middle Layer [Web] Services Higher Layer

Layering mandatory to address complexity

slide-5
SLIDE 5

5

OASIS WS-* Layering

Assurances Messaging SOAP WS-Security MTOM WS-Addressing Metadata WS-Policy WSDL WS-Discovery UDDI WS-Metadata Exchange WS-Transfer WS-Enumeration WS-Eventing XML Schema WS-Reliable Messaging WS-Coordination WS-Atomic Transaction WS-Business Activity WS-Trust WS-Secure Conversation

Profiles and

Infrastructure WS-Management WS-Federation Devices Profile Foundation SOAP / HTTP SOAP / UDP MIME XML Infoset XML 1.0 XML Namespaces

Layering in ‘Run time’ stack and Federation

Requestor Service Provider

Level 1

Requestor Service Provider

Level 2

Requestor

Level 3

Requestor

Level 4

Requestor

Level 5

Service Provider Service Provider Service Provider

Discovery Svc 1 Discovery Svc 2 Directory 1 (A.D.) Directory 2 (UDDI) I.D. Space 1 I.D. Space 2

  • Ent. Svc Bus 1
  • Ent. Svc Bus 2

TCP/IP Network

SOAP/SAML SOAP/SAML Workflow Mgr 1 Workflow Mgr 2

slide-6
SLIDE 6

6

Conceptual Model for Federation

Mappable Attributes Common Attributes Non Mappable Attributes

Is there a single model for Federation

  • Highly unlikely
  • Different models will be needed for

– Directory Federation [AD, UDDI, Relational Data Base] – Identity Federation

  • Identity Space Integration, ID attributes,

– SAML / Soap

  • Middleware specific messaging

– Enterprise Service Buses – Name spaces, Cross enterprise Bridging – Underlying TCP / IP Networking

slide-7
SLIDE 7

7

Data Transparency and Federation Data Transparency-- Schema Mapping

IBM Tool for mapping across schemas

slide-8
SLIDE 8

8

Data Transparency - Attribute Mapping

Data Transparency – Query Transformation

  • Query Transformation IBM Tool
slide-9
SLIDE 9

9

Directories Identities and Attribute Federation

Directories Background

  • Directory types considered for use are LDAP and x.500

– Based on RFCs – inetOrgPerson object class used for people – Based on commercial requirements

  • Active Directory

– User object class used for people – AD User object has inetOrgPerson attributes

  • DADIWG AD schema guidance for:

– Global address list attributes (people)

  • DMS provides x.500 schema guidance

– x.500 not included here

slide-10
SLIDE 10

10

Directory Scope and what it will do

  • Capabilities

– The objective is to implement a standard directory schema in accordance with DoDD 8100.1 that implicitly mandates the use of the Lightweight Directory Access Protocol (LDAP) for digital identities, resulting in a more efficient identity related data synchronization communications for the Air Force and Joint environment.

  • Directory ought to address

– Directory Information Tree (DIT) structure – People – Roles – Devices – Services [Middleware and application specific] – Object class and attribute naming conventions

  • Directory operations need to support:
  • Garrison
  • Tactical
  • Federation with external organizations

– LDAP and AD instantiations – UDDI

Directory Information Tree (1 of 2)

c=US

  • =U.S. Government
  • u=DoD
  • u=<Agency>
  • u=DoD Agencies
  • u=USA
  • u=USAF
  • u=USMC
  • u=USN
  • u=NOAA
  • u=USPHS
  • u=USCG
  • u=Affiliates
slide-11
SLIDE 11

11

Directory Information Tree (2 of 2)

c=US

  • =U.S.

Government

  • u=DoD
  • u=<Agency>
  • u=DoD

Agencies

  • u=USA
  • u=USAF
  • u-=USMC
  • u=USN
  • u=Devices
  • u=NOAA
  • u=USPHS
  • u=USCG
  • u=Affiliates
  • u=People
  • u=Roles
  • u=Services
  • u=A

Active Duty

  • u=B

Presidential appt

  • u=C

Civil Service

  • u=D

Disabled Veteran

  • u=Functional
  • u=Occupational
  • u=Operational
  • u=Organizational
  • u=E

Contractor

  • u=F

Former RR

  • r SR
  • u=I

Non-DoD Civ Svc

  • u=J

Academy Student

  • u=K

NAF

  • u=L

Lighthouse Srvc

  • u=M

Non-Govt Agency

  • u=N

National Guard

  • u=O

Non-DoD ctr

  • u=Q

Reserve retiree

  • u=R

Retired

  • u=T

Foreign military

  • u=U

Foreign national employee

  • u=V

Reserve

LDAP People Schema

  • Standard LDAP People Object Class

– inetOrgPerson represents people who are associated with an organization in some way. It is a structural class and is derived from the organizationalPerson class which is defined in X.521.

  • New Object Class

– dodNetOrgPerson is a auxiliary object class that is intended to hold attributes about people in or associated with the Department of Defense. – Derived from inetOrgPerson

slide-12
SLIDE 12

12

Active Directory People Schema

  • User People Object Class

– User represents people who are associated with an

  • rganization in some way. It is a structural class and

is derived from the organizationalPerson class which is defined in X.521.

  • New Object Class

– dodUserOrgPersonis a auxiliary object class that is intended to hold attributes about people in or associated with the Department of Defense. – Derived from inetOrgPerson

Unique Identifier for People

  • Attribute Name

– gigID

– Global Information Grid Identification

  • Format

– The DMDC assigned Electronic Data Interchange Person Identifier appended with the Personnel Category Code – [EDI-PI][PCC]. – Example “0123456789A”.

slide-13
SLIDE 13

13

Directory and Federation Issues

  • What directories and when?

– Do we use lowest common denominator and ignore richness? – Where are services and devices registered? – Possibly need to separate infrastructure and application spaces

  • Naming guidelines and relation to standards

– DOD Directive [draft] 8130 status? – Naming for devices, services, [sensors?]

  • What schemas and what are the models for schema

mapping?

  • What are the models for attribute mapping

– Common Attributes, Similar and Dissimiliar attributes – Domain specific attributes

  • Query Transformation across directories and domains
  • Who will take it to standards / consortia and get it accepted?
  • Scale up, Robustness and other issues

Naming and Federation Issues

  • Naming is fundamental to practically everything
  • Many different kinds of names being used wout integration
  • How does one build federated name spaces?
  • Unique Identifiers needed for Services, Systems, Objects,

Devices and Containers

– Must be globally unique – Root OIDs issued by ANSI

  • http://www.ansi.org/other_services/registration_programs/reg_org.aspx?me
  • ASD Initiative [DOD Draft Directive 8130]
  • Common Name

– [DNS prefix]-[Acronym]-[Description] – Example af-mil-AIMNT-Connection-Point

  • LDAP Display Name

– [DNS prefix]-[Acronym Description] – Example afmil-AIMNTConnectionPoint

slide-14
SLIDE 14

14

Authorization and Federation Issues

  • What pieces of information will be used for authorization

– What are the authoritative sources and how will provisioning take place? – ABACS helps somewhat but does not solve the problem – For ‘groups’ what are the definitions for each forest and where do cross forest or cross enterprise mappings take place?

  • WS-federation does not address these aspects
  • How are group semantics to be matched?

– How and who will build credentials to contain group information on a per forest and on a per invocation basis?

  • Requestor may select Groups A and B for invocation 1 but only Group B for

invocation 2

– How will revocation work? – What will be the relationship between COIs and Groups? – Questions similar to the above but now with ‘roles’

slide-15
SLIDE 15

15

Roles

  • Roles provide a mechanism to group identities that

have a common relationship.

  • There are several common relationships that support

grouping the types of roles into separate directory branches.

  • The intent is to provide a consistent methodology of

mapping users under the people branch with roles.

  • Role Based Access Control (RBAC) basis for Web

Standards (XCAML)

  • Standard role schema will support assigning

permissions in a more consistent manner between

  • perational directory implementations.

Role Types

Consists of a branch for job functions within the context of an organization with some associated semantics regarding the authority and responsibility conferred on individuals assigned to the role.

  • u=Organizational

Consists of a branch for the job categories associated with operational position or function.

  • u=Operational

Consists of a branch for the set of standard job categories that represent competency in a functional area.

  • u=Occupational

Consists of a branch for business functions in relationship to mission-applications.

  • u=Functional

Level Six Branch of Military Service, ou=Roles

slide-16
SLIDE 16

16

Major Issues across all Models

  • Who will define solutions and get it accepted across

services and ‘Joint’?

– What time frame – What is the interim guidance

  • We need things NOW!
  • Who will relate the new capabilities to commercially

available types?

– Policy based ‘management’ may not happen for a long time due to many challenges

  • Who will take the new solutions to consortia and standards

bodies and get the solutions accepted?

  • How and when will we know that other ‘ities’ are met?

– Scalability, Dependability, Interoperability …

Way Forward

  • Multiple coordinated activities

– AF, IDA, DISA, selected researchers and a few SMEs will develop and document simple federation models to serve as a start and provide recommendations akin to some level of program guidance for ‘programs’ in infancy

  • Results in 120 to 150 days

– IDA, Cornell, Berkeley with OSD /DISA and AFRL support will conduct sustained research and produce more detailed results with feasibility demonstrations based on extensions to vendor capabilities

  • Take results with DISA/ Services to consortia / standards …
  • Duration 24 months [staged results from 12 to 24 months]
  • DISA will set up ‘drum beat’ and plan / organize all major

planning activities and make decisions pertaining to governance, candidate selection, usage and acquisition [as needed]

– DISA will also orchestrate integration of federation capabilities with NCES and NCID

slide-17
SLIDE 17

17

Backups